Too many politicians spoiling the babu’s broth
A popular maxim in talking about governance in India is that “politicians come and go but it’s the Indian Administrative Service (IAS), or the bureaucrat, that runs the country”. There is perhaps a great deal of truth to this. One would be hard-pressed to think of government programmes that go from planning to execution without the involvement of bureaucrats.
There is, however, an issue with this arrangement: Politicians come and go but bureaucrats only come and rarely go. We can vote out incompetent politicians but we don’t have a way to incentivise bureaucrats to work well.
And incentives matter, unfortunately. So, do we just bite the bullet and endure the Sir Humphreys of the subcontinent? The short answer is, not necessarily. As recent research shows, the link between electoral consequences and bureaucratic performance is subtle. Politicians can and do provide incentives to bureaucrats if the stars are properly aligned.
No, I am not talking about any cosmic interference here but there is indeed an element of luck involved. That element is how administrative and electoral boundaries are aligned. In India, the areas of jurisdiction under a bureaucrat and a politician often do not coincide. That is, an area under the jurisdiction of a bureaucrat does not entirely belong to a single politician’s constituency.
This observation lets us study the behaviour of a bureaucrat with multiple politicians to understand if the same bureaucrat deals with different politicians differently. Two recent papers adopt this approach to show that when a bureaucrat reports to more than one politician, performance is compromised.
In Politicians, Bureaucrats And Development: Evidence From India, Saad Gulzar from New York University and Benjamin Pasquale, an independent researcher, study the differences in implementation of the Mahatma Gandhi National Rural Employment Guarantee Act (MNREGA) when a bureaucrat reports to a single politician versus multiple politicians. They find that when a bureaucrat reports to multiple politicians, monitoring is under-provided—resulting in poor execution, reflected by fewer workdays and fewer workers.
Interestingly, they also find that if the two politicians belong to different parties, the outcomes are worse than when they are from the same party. Moreover, as the number of politicians a bureaucrat reports to increases, the performance worsens.
The main takeaway is that when politicians internalize the electoral benefits from a bureaucrat’s performance—wherein the former can claim credit for better implementation—bureaucrats perform significantly better. The authors carefully rule out a number of other mechanisms behind this finding to deliver a succinct message: “Political influence may be more favourable to development than is commonly assumed.”
A working paper by Anusha Nath of the Federal Reserve Bank of Minneapolis, Bureaucrats And Politicians: Electoral Consequences And Dynamic Incentives, studies the effect of electoral competition on bureaucratic performance. We often assume that more competition is better but Nath shows that more competition can have perverse consequences.
Imagine a bureaucrat nearing his promotion cycle. Say, there are two politicians: a stronghold one with a high chance of re-election and a non-stronghold one. If the promotion falls before the election, the bureaucrat must please both types of politicians to earn good reviews but if the promotion happens after the election, the bureaucrat will strategically please only the politician who has been re-elected.
Nath looks at the time taken to approve projects and finds a systematic difference between the way a bureaucrat performs with a stronghold politician versus the way the bureaucrat performs with a non-stronghold one.
Clearly, too much political competition,which lowers the probability of re-election, results in poor performance by bureaucrats. In fact, this may engender a vicious cycle whereby citizens punish the politicians for underperforming bureaucrats.
But we cannot possibly argue that since bureaucratic performance is positively correlated with re-election probability, we should get rid of elections. Neither can we imagine a round of elections for bureaucrats. So, where does this leave us? How can we use the existing political institutions to provide better incentives to bureaucrats?
The key question to ask here is: what is the source of this inefficiency? It is the distinction between the political and bureaucratic boundaries. These boundaries affect the incentive structure facing both the politicians and the bureaucrats and can have serious implications for the services provided to citizens.
Could we rework one of the boundaries so that we minimize the non-overlap as much as possible?
This could involve some short-run costs but possibly the long-run benefits could be worth it. Since the delivery of most services relies on bureaucrats, it behoves us to seriously try and get the right incentive structure for bureaucrats who otherwise have no electoral accountability. Given where we stand in the pyramid of development, simply picking such low-hanging fruits can get us very far.
Aditya Kuvalekar is a PhD candidate at New York University.
Comments are welcome at email@example.com